动车运行驾驶员头部姿态图像特征识别仿真  

Simulation of Recognition of the Vehicle Driver's Head Posture Image Feature

作　　者：梁利娟[1] 阮莹[1] 

机构地区：[1]河南师范大学新联学院,河南郑州451400

出　　处：《计算机仿真》2018年第1期374-377,共4页Computer Simulation

摘　　要：对动车运行驾驶员头部姿态图像的特征进行识别,能够有效提高动车运行时驾驶员的安全系数。对头部姿态图像特征的识别,需要建立头部旋转姿态模型,估计头部姿态,完成头部姿态图像特征识别。传统方法首先获得全局姿态标签,定位特征点,但忽略了估计头部姿态,导致识别精度低。提出鼻子眼睛作为定位点进行动车运行驾驶员头部姿态图像特征识别方法。定位特征点以适应大角度偏转,计算旋转角度,建立头部旋转姿态模型,估计头部姿态,实现动车运行驾驶员头部姿态图像特征识别。实验结果表明,hhsy方法有效提高了动车运行时驾驶员的安全系数。ABSTRACT ： The recognition of characteristics of posture image of head can effectively improve the safety factor of driver when motor train is running. In order to recognize the feature of head pose, it is necessary to establish posture model of head and estimate head pose, so as to complete the feature recognition of head pose image. Traditionally, some methods ignore the estimation of head pose, resulting in low recognition accuracy. ＂A new method is proposed for recognizing the features of head posture image of motor train in running when the nose and eyes are used as positioning points. The feature points are adapted to wide-angle deflection. The angle of rotation is calculated to build rotation posture model of head. Moreover, the head pose is estimated. Thus, the feature recognition of head pose of motor train is realized. Simulation results show that the proposed method can effectively improve the safety factor of driver when motor train is running.

关 键 词：动车运行 头部姿态图像 特征识别 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]